Index  | Recent Threads  | Unanswered Threads  | Who's Active  | Guidelines  | Search
 
World Community Grid Forums
  Category: Beta Testing
  Forum: Beta Test Support Forum
  Thread: New Beta Test – May 29, 2019 [Issues Thread]

Quick Go »
No member browsing this thread
Thread Status: Active
Total posts in this thread: 315
Posts: 315   Pages: 32   [ Previous Page | 1 2 3 4 5 6 7 8 9 10 | Next Page ]
[ Jump to Last Post ]
Post new Thread
Author
Previous Thread This topic has been viewed 58176 times and has 314 replies Next Thread
Dark Angel
Veteran Cruncher
Australia
Joined: Nov 11, 2005
Post Count: 721
Status: Offline
Project Badges:
Reply to this Post  Reply with Quote 
Re: New Beta Test – May 29, 2019 [Issues Thread]
Ok, all three done and uploaded.
Two are reporting a fraction over 14 hours. That's bollocks, they both went over 18 hours.

Result Name Device Name Status Sent Time Time Due /
Return Time CPU Time / Elapsed Time (hours) Claimed/ Granted BOINC Credit
BETA_ ARP1_ 0001780_ 000_ 1-- Mint-Boxen Pending Validation 5/29/19 18:08:09 5/30/19 09:30:46 14.10 / 14.14 31.7 / 0.0
BETA_ ARP1_ 0000302_ 000_ 0-- Mint-Boxen Pending Validation 5/29/19 17:45:54 5/30/19 07:52:54 14.04 / 14.07 31.7 / 0.0
BETA_ ARP1_ 0000031_ 000_ 0-- Tall-Boxen Pending Validation 5/29/19 13:34:40 5/30/19 10:07:07 19.45 / 19.45 31.7 / 0.0

I believe the dog loves them. ;o)

Bollocks?
Been so long since he had any I doubt he remembers them.
----------------------------------------

Currently being moderated under false pretences
[May 30, 2019 12:13:00 PM]   Link   Report threatening or abusive post: please login first  Go to top 
Jean-David Beyer
Senior Cruncher
USA
Joined: Oct 2, 2007
Post Count: 335
Status: Offline
Project Badges:
Reply to this Post  Reply with Quote 
Re: New Beta Test – May 29, 2019 [Issues Thread]
I got four of these work units yesterday, projected to run 2 hours 59 minutes each; deadline Wednesday June 5. None are running yet; three HSTB are running and one FAH2 is running. Linux OS.
----------------------------------------
[May 30, 2019 12:52:22 PM]   Link   Report threatening or abusive post: please login first  Go to top 
pvh513
Senior Cruncher
Joined: Feb 26, 2011
Post Count: 260
Status: Offline
Project Badges:
Reply to this Post  Reply with Quote 
Re: New Beta Test – May 29, 2019 [Issues Thread]
Here are some more reliable stats on my WUs

65.145% complete in 16h05 on Xeon E5-2630 v3 with HT
62.541% complete in 15h32 on EPYC 7281 with SMT
66.729% complete in 15h10
67.500% complete in 15h06
68.812% complete in 15h04 all on Ryzen 7 1700X with SMT

disk usage has risen to about 650 MiB per slot. Just did another cycle of suspend/resume with LAIM off and saw nothing unusual. I will now let them run to finish.

Edit: forgot the OS: Linux openSUSE Leap 42.3 / 15.0
----------------------------------------
[Edit 1 times, last edit by pvh513 at May 30, 2019 1:09:47 PM]
[May 30, 2019 12:59:51 PM]   Link   Report threatening or abusive post: please login first  Go to top 
Former Member
Cruncher
Joined: May 22, 2018
Post Count: 0
Status: Offline
Reply to this Post  Reply with Quote 
Re: New Beta Test – May 29, 2019 [Issues Thread]
Reported first of three.
Seems it thinks it only took 14.07 minutes to run.
It was more like 18 hours.
and 32 credits for 18 hours work is garbage.

Restarted clients with remaining two.
One picked up fine, the other seems to have lost about 7 hours of run time. Neither errored out as yet.

Option 1) Though the announcement gave an indication of the runtime, they did not pack an appropriate matching norm flops in the header, and thus all are deemed slugs.
Option 2) Dr. Anderson gets royalties on the points earned, and thus they decided to do a Donald J. Trump number on him. He anyway thinks IBM is wasting time on climate, sorry, weather science.
[May 30, 2019 1:02:45 PM]   Link   Report threatening or abusive post: please login first  Go to top 
KerSamson
Master Cruncher
Switzerland
Joined: Jan 29, 2007
Post Count: 1671
Status: Offline
Project Badges:
Reply to this Post  Reply with Quote 
Re: New Beta Test – May 29, 2019 [Issues Thread]
First reported of 4 computed by a i7-6700K @4 GHz, no OC, 16 GB RAM, LinuxMint 19.1 x64 (kernel 4.15.0-50-generic) with Cinnamon:
BETA_ARP1_0000106_000	5/29/19 17:40:12	5/30/19 10:40:56	14.78 / 14.79	583.7 / 0.0

Very few restore points, 2+ hours between to RP.
Used memory by the WU: about 700 MB
The host hung after a user login and logout and required a reboot to continue computing the beta WU.
The claimed credit is low: about 39 points / hour comparing to an average of 67 points / hour with Zika.
During the first 10 hours, the estimated computation time was about 50 min. After 10 hours, the estimated computation time was more realistic.
I will report further information as soon as the other Beta WUs are completed.

Update #01 - 2019-05-30 23:45 (UTC)
Very few restore points, 2+ hours between to RP.
BETA_ARP1_0000106_000	H#1 	5/29/19 17:40:12	5/30/19 10:40:56	14.78 / 14.79	583.7 / 0.0
BETA_ARP1_0001826_000	H#2	5/29/19 18:09:39	5/30/19 20:57:40	25.35 / 25.39	31.7 / 0.0
BETA_ARP1_0001881_000	H#2	5/29/19 18:09:39	5/30/19 20:57:40	25.77 / 25.81	31.7 / 0.0
BETA_ARP1_0001666_000	H#3	5/29/19 18:07:00	5/30/19 21:06:58	15.79 / 15.82	31.7 / 0.0

  • H#1
    i7-6700K @4 GHz, no OC, 16 GB RAM, LinuxMint 19.1 x64 (kernel 4.15.0-50-generic) with Cinnamon, Boinc v7.6.31
    Used memory by the WU: about 700 MB
    The claimed credit is low: about 39 points / hour comparing to an average of 67 points / hour with Zika
    During the first 10 hours, the estimated computation time was about 50 min. After 10 hours, the estimated computation time was more realistic.
    Beside the Beta WU, 7 Zika WUs were concurrently computed.
  • H#2
    Phenom II x6 @2.8 GHz, no OC, 8 GB RAM, Ubuntu 14.04 x64 (kernel 4.4.0-148-generic) with Unity, Boinc v7.2.42
    Used memory by the WU: about 685 MB
    The claimed credit is ridiculous: about 1.25 points / hour comparing to an average of 49 points / hour with Zika
    Estimated computation time relatively realistic.
    Beside the 2 Beta WUs, 4 Zika WUs were concurrently computed.
  • H#3
    Ryzen 2700 @3.2 GHz, no OC, 16 GB RAM, LinuxMint 19.1 x64 (kernel 4.15.0-50-generic|libc 2.27) with Cinnamon, Boinc v7.9.3
    The claimed credit is ridiculous: about 2 points / hour comparing to an average of 56 points / hour with Zika
    Beside the Beta WU, 13 Zika WUs and 2 MIP1 WUs were concurrently computed, without negative impact on the computation duration for the MIP1 WUs.

Host H#1 hung after a user login and logout and required a reboot to continue computing the beta WU.

Update #02 - 2019-05-31 11:50 (UTC)
Update #03 - 2019-06-01 07:20 (UTC)
Update #04 - 2019-06-04 08:40 (UTC)
BETA_ARP1_0001826_000	H#2	5/29/19 18:09:39	5/30/19 20:57:40	25.35 / 25.39	31.7 / 31.7 sad
BETA_ARP1_0001881_000	H#2	5/29/19 18:09:39	5/30/19 20:57:40	25.77 / 25.81	31.7 / 31.7 sad
BETA_ARP1_0001666_000	H#3	5/29/19 18:07:00	5/30/19 21:06:58	15.79 / 15.82	31.7 / 31.7 sad
BETA_ARP1_0000106_000	H#1 	5/29/19 17:40:12	5/30/19 10:40:56	14.78 / 14.79	583.7 / 307.7

1.25 points / hour !!! (H#2) resp. 2 points / hour !!! (H#3)
20.8 points / hour !!! (H#1) i.e about 30% of the daily average with Zika.

!!! For further updates, see post: https://www.worldcommunitygrid.org/forums/wcg/viewpostinthread?post=611204
Cheers,
Yves
---
Beside the Beta WU, 7 Zika WUs were computed by the other threads.
----------------------------------------

Décrypthon team progress - KerSamson's contribution

----------------------------------------
[Edit 13 times, last edit by KerSamson at Jun 9, 2019 8:50:50 AM]
[May 30, 2019 1:30:31 PM]   Link   Report threatening or abusive post: please login first  Go to top 
Former Member
Cruncher
Joined: May 22, 2018
Post Count: 0
Status: Offline
Reply to this Post  Reply with Quote 
Re: New Beta Test – May 29, 2019 [Issues Thread]
Is this project leeching cycles from other sciences or is it a coincidence that the concurrently running MIP1 are now needing 9 hours CPU time instead of the average 3 they normally take? ZIKA and HST1 seem not to be showing a change. Efficiencies are unchanged.
----------------------------------------
[Edit 1 times, last edit by Former Member at May 30, 2019 2:20:44 PM]
[May 30, 2019 2:20:24 PM]   Link   Report threatening or abusive post: please login first  Go to top 
Crystal Pellet
Veteran Cruncher
Joined: May 21, 2008
Post Count: 1316
Status: Offline
Project Badges:
Reply to this Post  Reply with Quote 
Re: New Beta Test – May 29, 2019 [Issues Thread]
Each individual work unit represents 48 hours calendar time for this simulation.

The 12.5% checkpoint interval correlates with the 48 hours and the interval of 6 hours
between the meteorological input data, usual GRIB-data up to 32 atmospheric levels.
One task is processing the data for a small area for 8 timestamps. E.g.
20180401000000 UTC (YYYYMMDDHHMMSS)
20180401060000 UTC
20180401120000 UTC
20180401180000 UTC
20180402000000 UTC
20180402060000 UTC
20180402120000 UTC
20180402180000 UTC
----------------------------------------
[May 30, 2019 2:28:51 PM]   Link   Report threatening or abusive post: please login first  Go to top 
Seoulpowergrid
Veteran Cruncher
Joined: Apr 12, 2013
Post Count: 815
Status: Offline
Project Badges:
Reply to this Post  Reply with Quote 
Re: New Beta Test – May 29, 2019 [Issues Thread]
Running some of a Ryzen Windows desktop. They are around 20 hours runtime in, each is using nearly 500 megs of RAM per with one odd one in the group taking 310 megs. No issues thus far.
----------------------------------------
[May 30, 2019 4:17:11 PM]   Link   Report threatening or abusive post: please login first  Go to top 
Former Member
Cruncher
Joined: May 22, 2018
Post Count: 0
Status: Offline
Reply to this Post  Reply with Quote 
Re: New Beta Test – May 29, 2019 [Issues Thread]
@lavaflow

Others have (speculated?) said that you shouldn't run too many copies of MIP at once because it is "heavy" on certain aspects of CPU/system design (local cache?). Since this beta is heavy on memory, it may also be heavy on the same aspects of your system. I'm a bit surprised at your comments that efficiencies are not affected though ...

I wonder how many beta+MIP you are running in parallel, compared to how many MIP you normally run in parallel? On my system I only ran one MIP with a single beta, but don't see a problem running two or even three MIP in parallel normally. I saw nothing similar to what you're seeing on this occasion.

Of course, it could just be normal task run-time variation ...
[May 30, 2019 4:58:28 PM]   Link   Report threatening or abusive post: please login first  Go to top 
Former Member
Cruncher
Joined: May 22, 2018
Post Count: 0
Status: Offline
Reply to this Post  Reply with Quote 
Re: New Beta Test – May 29, 2019 [Issues Thread]
I'm aware which is why app_config allows not more than 3 running which they were. They ended though just over 6 hours, not, nine, but still extraordinary long, double the running average, and scanning my database, none have so far on this machine.

edit, example

7.16 MIP1_00193819_0659_0 07:45:41 (06:17:44) 5/30/2019 7:17:33 PM 5/30/2019 7:17:58 PM 81.11 Reported: OK 270.43 MB 230.90 MB

Daytime efficiency BTW, at night running 96-98%
----------------------------------------
[Edit 2 times, last edit by Former Member at May 30, 2019 6:28:18 PM]
[May 30, 2019 6:26:03 PM]   Link   Report threatening or abusive post: please login first  Go to top 
Posts: 315   Pages: 32   [ Previous Page | 1 2 3 4 5 6 7 8 9 10 | Next Page ]
[ Jump to Last Post ]
Post new Thread